Re: Why partition

[Albert DE WINT <adwint pandora be>]

Op di 17-06-2003, om 22:33 schreef Rick Stevens:
> Mark Knecht wrote:
> > On Tue, 2003-06-17 at 10:26, Otto Haliburton wrote:
> > 
> >>Anyway you store data on a
> >>disk creates problems, but partitioning allows optional uses for the
> >>same disk by different OS's.
> >>
> > 
> > 
> > Or applications...
> > 
> > In my audio recording, where I am likely to record many files that are
> > many minutes long, stereo 24-bit, 44.1KHz audio uses about 16MB/Min. A
> > four minute song uses 64MB/stereo track, and I'm generally dealing with
> > 20 of these.
> > 
> > On my audio partitions I like 32K or 64k allocation as it significantly
> > reduces the number of interrupts generated by the hardware and moves the
> > data more efficiently. Wasting 31K at the end of a track is no big deal
> > from a storage point of view.
> > 
> > On my program partitions I use 4K allocations.
> > 
> > Just another 2 cents... ;-)
> 
> In for $.02, in for a buck :-P
> 
> Changing your allocation size won't reduce the number of interrupts
> unless:
> 
> 	a) the allocation size divided by the physical sector size (512
> 	bytes on 95% of hard drives, 2048 bytes on CD-ROM) is evenly
> 	divisible into the sectors-per-cylinder of the drive since the
> 	interrupts are generally caused by head movement from cylinder
> 	to cylinder; and
> 
> 	b) the thing was set up on a fresh, clean, filesystem so that
> 	the starting point is aligned (normal, since there are very few
> 	filesystems that would let you change the allocation size after
> 	setup)
> 
> I think the gist of this thread is that in the real world, partitioning
> is mandatory to keep the disk maps to a reasonable size and keep access
> speeds up.  It may not be the most efficient use of the space, but the
> vast majority of computing is time critial, not storage critical.
> 
> The problem with theories is that they often don't concern themselves
> with real-world requirements.  If your cat has fleas, theoretically the
> most effective treatment is to toss the cat into the incinerator
> (intense heat is a fail-safe method of killing fleas).  The problem is,
> it kills cats, too, but the theory doesn't take that into account.

Yep, though my cat would probably pass the fleas to me, before she'd
ever make it to the incinerator. That's my cat's idea about partitioning
fleas...
> 
> I think we've flogged this thread enough.  I could be wrong, though. ;-)
> ----------------------------------------------------------------------
> - Rick Stevens, Senior Systems Engineer     rstevens vitalstream com -
> - VitalStream, Inc.                       http://www.vitalstream.com -
> -                                                                    -
> -             Polygon: A dead parrot (sorry, John Cleese!)           -
> ----------------------------------------------------------------------
> 
> 
> _______________________________________________
> Redhat-install-list mailing list
> Redhat-install-list redhat com
> https://www.redhat.com/mailman/listinfo/redhat-install-list
> To Unsubscribe Go To ABOVE URL or send a message to:
> redhat-install-list-request redhat com
> Subject: unsubscribe
> 
>